movement: Modelling and Analysing Human Movement Data

gravity.movement

R Documentation

Use the gravity model to predict movements across a landscape.

Description

Given a (dense) distance matrix distance giving the euclidean distances beween all pairs of sites, a vector of population sizes at these sites population, and a set of parameters theta, use the function flux to predict movement between all sites using a version of the gravity model (Viboud et al. 2006).

The model can be sped up by setting minpop and maxrange, see gravity.flux for details. If progress = TRUE a text progress bar is displayed and starting and total times reported.

Note that this is a pretty shitty prototype at the moment and will be re-written in a much more efficient way in the future!

Usage

  gravity.movement(distance, population, theta, minpop = 0, maxrange = Inf, progress = TRUE)

Arguments

`distance`	a distance matrix giving the euclidean distance between pairs of sites
`population`	a vector giving the population at all sites
`theta`	a vector of parameters in the order scalar, exponent on donor pop, exponent on recipient pop, exonent on distance
`minpop`	the minimum population size to consider (by default 1, consider all sites)
`maxrange`	the maximum distance between sites to consider (by default `Inf`, consider all sites)
`progress`	whether to display a progress bar and start and end times - can be useful for big model runs

Value

A (dense) symmetric matrix giving predicted movements between all sites

Examples


# generate random coordinates and populations
n <- 30
coords <- matrix(runif(n * 2), ncol = 2)
pop <- round(runif(n) * 1000)

# calculate the distance between pairs of sites
d <- as.matrix(dist(coords))

# predict movement between them using the radiation model
move <- gravity.movement(d, pop,theta=c(1e-4,0.6,0.3,3))

# plot the points
plot(coords, pch = 16, cex = pop / 500,
     col = 'grey40', axes = FALSE,
     xlab = '', ylab = '')

# and add arrows showing movement
for(i in 2:n) {
  for(j in  (i - 1):n) {
    arrows(coords[i, 1],
           coords[i, 2],
           coords[j, 1],
           coords[j, 2],
           lwd = 2,
           length = 0.1,
           col = rgb(0, 0, 1, move[i, j] / (max(move) + 1)))
  }
}

SEEG-Oxford/movement documentation built on April 17, 2023, 4:17 p.m.